Last Week’s AI News #31

For a long time, AI felt like something companies were racing to build.A competition of models, benchmarks, funding rounds, and product launches.

Who has the best model?
Who is ahead on performance?
Who ships faster?

That was the focus.

But that phase is starting to shift.

Because the real impact of AI is no longer happening at the model level. It’s happening inside companies.

Inside workflows.
Inside teams.
Inside the way work actually gets done.

AI is no longer just a tool you choose to use. It’s becoming a layer that reshapes how organizations operate from within.

And the companies that understand this shift early will move very differently from those that don’t.

Here’s what you need to know from last week in AI:

• Google rallies DeepMind to out-code Anthropic
• Adobe launches CX Enterprise for agentic workflows
• OpenAI releases ChatGPT Images 2.0
• Meta tracks employee activity for AI training
• Google launches Deep Research agents
• OpenAI introduces Workspace Agents
• White House targets China over AI distillation
• Anthropic report shows rising anxiety among top AI users
• Everything else that happened in AI last week 

GOOGLE RALLIES DEEPMIND TO OUT-CODE ANTHROPIC

Google co-founder Sergey Brin is personally pushing DeepMind to close the gap with Anthropic in coding, forming a new internal “strike team” focused on Gemini.

The group is led by Sebastian Borgeaud under CTO Koray Kavukcuoglu, with Brin positioning coding as the fastest path toward self-improving AI systems. Internally, researchers reportedly rate Claude’s coding abilities higher than Gemini’s, which triggered the initiative.

Engineers are now required to use Google’s internal AI agents for complex tasks, with performance tracked through a leaderboard system.

Why does it matter for businesses?
This is not just about better models, it’s about automating internal operations. Google is trying to turn AI into a layer that improves itself and the company at the same time, a direction already visible inside competitors like OpenAI and Anthropic.

ADOBE LAUNCHES CX ENTERPRISE FOR AGENTIC WORKFLOWS

Adobe introduced CX Enterprise, a new platform designed to orchestrate marketing, content, and customer interactions through networks of AI agents.

The system combines brand visibility, content production, and customer engagement into a single orchestration layer. A “Coworker” agent dynamically assembles tools and agents to execute multi-step tasks based on user goals.

Adobe’s agents can also integrate with external systems like ChatGPT, Claude, and Gemini, while a new skills catalog enables reusable workflows across organizations.

Why does it matter for businesses?
Creative and marketing workflows are becoming fully agent-driven. But the bigger shift is happening at the model level, as labs build end-to-end systems, platforms like Adobe risk being bypassed entirely.

OPENAI RELEASES CHATGPT IMAGES 2.0

OpenAI launched ChatGPT Images 2.0, a major upgrade to its image generation system, calling it the most advanced model it has built.

The model introduces a “thinking” step before generating images, allowing it to plan, gather references, and validate outputs. It ranks first on image generation benchmarks and supports high-resolution outputs, multiple formats, and multilingual rendering.

Sam Altman described the leap as comparable to jumping multiple model generations at once.

Why does it matter for businesses?
This is not just better image quality, it changes how visual content is created. Planning and reasoning inside generation opens entirely new workflows, reducing iteration time and expanding creative possibilities.

META TRACKS EMPLOYEE ACTIVITY FOR AI TRAINING

Meta launched an internal initiative to record employee activity, including screenshots, keystrokes, and mouse movements, to train its AI systems.

The program focuses heavily on developers and logs activity across tools like coding environments and internal AI assistants. Employees reportedly have no option to opt out, and the rollout coincides with upcoming layoffs.

The company frames it as a way to improve models through real-world usage data.

Why does it matter for businesses?
This mirrors how robotics systems learn from human behavior, but applied to digital work. It signals a future where everyday workflows become training data, raising both efficiency opportunities and serious ethical questions.

GOOGLE LAUNCHES DEEP RESEARCH AGENTS

Google released Deep Research and Deep Research Max, advanced agents designed to generate full research reports using web data, files, and private data sources.

Built on Gemini 3.1 Pro, the agents can combine multiple data sources and produce structured outputs with charts and insights. Google is also partnering with data providers to integrate premium datasets directly into workflows.

The system turns complex research tasks into automated processes.

Why does it matter for businesses?
Research-heavy roles are becoming programmable. What used to take hours of analysis can now be integrated into products as a feature, significantly lowering the cost of high-level insights.

OPENAI INTRODUCES WORKSPACE AGENTS

OpenAI launched Workspace Agents in ChatGPT, enabling teams to deploy shared AI agents that handle multi-step workflows across tools like Slack.

Powered by Codex, these agents can retain memory, interact with connected systems, and operate autonomously on schedules. Teams can create and share custom agents with defined permissions and controls.

The feature builds on earlier GPTs but shifts toward collaborative, team-level usage.

Why does it matter for businesses?
Most companies already use AI in fragmented ways. Workspace agents aim to unify those efforts into structured systems, turning scattered experiments into scalable operations.

WHITE HOUSE TARGETS CHINA OVER AI DISTILLATION

The White House issued a memo accusing Chinese firms of large-scale AI distillation practices, using outputs from leading U.S. models to train their own systems.

Officials claim this is done through API abuse and jailbreak techniques, with potential policy responses including export restrictions. The move comes ahead of a major diplomatic meeting between the U.S. and China.

Chinese officials denied the claims, calling them unfounded.

Why does it matter for businesses?
AI competition is increasingly geopolitical. Regulations, restrictions, and global tensions will directly impact access to models, data, and markets.

ANTHROPIC REPORT SHOWS RISING ANXIETY AMONG TOP AI USERS

Anthropic published new research showing that workers who benefit most from AI are also the most concerned about job displacement.

The study links usage data with survey responses, revealing that high-frequency users, especially engineers, express significantly higher levels of concern. Early-career professionals are the most affected group.

While productivity gains are clear, they are often accompanied by increased workload and uncertainty.

Why does it matter for businesses?

AI adoption is creating a paradox: the more valuable the tools become, the more uncertainty they introduce. Companies need to manage not just productivity, but also employee perception and trust.

EVERYTHING ELSE THAT HAPPENED IN AI LAST WEEK

• OpenAI Chronicle: OpenAI introduced Chronicle, a feature using background agents to capture screen activity and build persistent memory. This signals a shift toward AI deeply embedded in daily workflows.
• Tinder and Zoom partnership with World: Tinder and Zoom are adding “proof of humanity” via iris scans to fight bots. This highlights the growing need to verify humans in AI-driven environments.
• Anthropic and Amazon expansion: Anthropic expanded its deal with Amazon, securing massive compute resources. This shows that advanced AI capabilities increasingly depend on infrastructure scale.
• Recursive Superintelligence funding: A new startup raised $500M to build self-improving AI systems. This signals a shift toward automating AI development itself.
• Core Automation launch: Former OpenAI and DeepMind researchers launched a lab focused on building AI that creates AI. This reflects accelerating innovation cycles in the field.
• Google DESIGN.md: Google open-sourced a file format that helps AI understand design systems and brand rules. This enables more automated and consistent creative workflows.
• Exa Deep Max: Exa launched a new agentic search tool that outperforms competitors in speed and accuracy. This shows search evolving into fully automated insight generation.
• Genspark Build: A new tool allows users to generate apps and websites directly from text prompts. This lowers technical barriers for businesses building digital products.
• ChatGPT for Clinicians: OpenAI released a tool for healthcare professionals that outperforms doctors in benchmarks. This shows AI rapidly entering high-skill industries.
• Meta layoffs driven by AI: Meta plans to cut 10% of its workforce citing AI efficiency gains. This signals that AI is already reshaping the job market.

AI is no longer just a race between companies. It’s becoming a system that reshapes how those companies operate internally.

Some will use it to move faster. Others will struggle to adapt their own structures.

But the shift is already happening.

👉 Subscribe to get next week’s AI before everyone else does.